3 ms·

SBF Was Reckless from the Start

46 points by feross 3y ago

ioblomov 3y ago
https://archive.ph/UARa8 https://archive.ph/UARa8
robocat 3y ago
https://archive.ph/JLwuQ https://archive.ph/JLwuQ There is a Lovely discussion on a self-referential bet made by SBF while working at Jane Street. Unfortunately the details matter a lot so I would love to hear from Asher's side. Matt writes plenty of interesting points on this bet in the footnotes. Especially interesting this comment on SBF and Kelly Criterion: https://qoppac.blogspot.com/2022/11/if-youre-so-smart-how-come-youre-not.html https://qoppac.blogspot.com/2022/11/if-youre-so-smart-how-co... One morning, another intern, called “Asher” in the book, offers to bet Bankman-Fried on the maximum amount any intern will lose that day. “This cannot settle above one hundred or below zero, right?” confirms Bankman-Fried. And then he buys at 65: If any intern loses more than $65 that day, Asher will pay Bankman-Fried that intern’s losses above $65; if none do, Bankman-Fried will pay Asher the difference between $65 and the maximum loss. If an intern loses $100, the maximum, then Bankman-Fried gets $35. If the biggest loser loses only $50, he pays Asher $15. Of course Bankman-Fried can easily win this bet by losing $100 himself.
- Vecr 3y ago
  How SBF is described as acting in this case as absolutely not rational if he was actually trying to maximize his expected number of dollars, there's going to be an optimal set of trades you can make with the finite number of dollars you start out with, and it's not going to be slamming your head against a single bet that's getting less and less attractive as you get deeper into the conditional probabilities. Make a single "good" bet (like SBF claimed to have done here) if you want, but use the rest of the money for less correlated bets that at least aren't strictly conditional on the first bet going well to have any utility at all.
  - sdwr 3y ago
    After letting the anecdote percolate for a bit, none of it makes sense. The idea that the other person is getting more and more humiliated, as their losses go from 98 dollars to 99 to 99.50 doesn't really track.
    - Vecr 3y ago
      Maybe loss of money is very closely linearly related to humiliation for SBF, but it is not for most people. That's the least confusing part of the story to me.
  - legalbeagle 3y ago
    This course of action (taking a small risk of losing everything in exchange for a high probability small reward) is what Nassim Taleb calls "picking up coins in front of a steamroller."
- robocat 3y ago
  Matt Levine updated the above: First, in the original version of yesterday’s column, I transposed the names in my first description of the bet, writing “If any intern loses more than $65 that day, Bankman-Fried will pay Asher that intern’s losses above $65; if none do, Asher will pay Bankman-Fried the difference between $65 and the maximum loss,” where I meant “If any intern loses more than $65 that day, Asher will pay Bankman-Fried that intern’s losses above $65; if none do, Bankman-Fried will pay Asher the difference between $65 and the maximum loss.” This is now corrected on the web and the Bloomberg Terminal, but if you read Money Stuff by email you got the wrong description of the bet, and that is annoying and I’m sorry. Second, a reader tweeted a better trade: “SBF also either missed or passed on a one-sided bet with greater +EV that would have inflicted greater loss on the counterparty on the first flip: pay two other interns $1 each to coin flip each other for $100.” Yes. This trade would cost Asher $35 (one intern loses $100 on the coin flip, $35 more than the $65 strike price on Bankman-Fried’s bet with Asher[3] ) and would make Bankman-Fried $33 ($35 from Asher minus the $2 he pays the interns). Crucially, Bankman-Fried’s actual trade — pay an intern $1 to flip a coin with him for $98 — had a slightly lower $32 expected value, and was risky; he had a 50/50 chance of winning $130 or losing $66. This trade — pay two interns to flip each other — has a guaranteed profit of $33 with no risk at all. Bankman-Fried went with the riskier trade.

46 points by feross 3y ago

https://archive.ph/UARa8 https://archive.ph/UARa8

https://archive.ph/JLwuQ https://archive.ph/JLwuQ There is a Lovely discussion on a self-referential bet made by SBF while working at Jane Street. Unfortunately the details matter a lot so I would love to hear from Asher's side. Matt writes plenty of interesting points on this bet in the footnotes. Especially interesting this comment on SBF and Kelly Criterion: https://qoppac.blogspot.com/2022/11/if-youre-so-smart-how-come-youre-not.html https://qoppac.blogspot.com/2022/11/if-youre-so-smart-how-co... One morning, another intern, called “Asher” in the book, offers to bet Bankman-Fried on the maximum amount any intern will lose that day. “This cannot settle above one hundred or below zero, right?” confirms Bankman-Fried. And then he buys at 65: If any intern loses more than $65 that day, Asher will pay Bankman-Fried that intern’s losses above $65; if none do, Bankman-Fried will pay Asher the difference between $65 and the maximum loss. If an intern loses $100, the maximum, then Bankman-Fried gets $35. If the biggest loser loses only $50, he pays Asher $15. Of course Bankman-Fried can easily win this bet by losing $100 himself.

Vecr 3y ago
How SBF is described as acting in this case as absolutely not rational if he was actually trying to maximize his expected number of dollars, there's going to be an optimal set of trades you can make with the finite number of dollars you start out with, and it's not going to be slamming your head against a single bet that's getting less and less attractive as you get deeper into the conditional probabilities. Make a single "good" bet (like SBF claimed to have done here) if you want, but use the rest of the money for less correlated bets that at least aren't strictly conditional on the first bet going well to have any utility at all.
- sdwr 3y ago
  After letting the anecdote percolate for a bit, none of it makes sense. The idea that the other person is getting more and more humiliated, as their losses go from 98 dollars to 99 to 99.50 doesn't really track.
  - Vecr 3y ago
    Maybe loss of money is very closely linearly related to humiliation for SBF, but it is not for most people. That's the least confusing part of the story to me.
- legalbeagle 3y ago
  This course of action (taking a small risk of losing everything in exchange for a high probability small reward) is what Nassim Taleb calls "picking up coins in front of a steamroller."
robocat 3y ago
Matt Levine updated the above: First, in the original version of yesterday’s column, I transposed the names in my first description of the bet, writing “If any intern loses more than $65 that day, Bankman-Fried will pay Asher that intern’s losses above $65; if none do, Asher will pay Bankman-Fried the difference between $65 and the maximum loss,” where I meant “If any intern loses more than $65 that day, Asher will pay Bankman-Fried that intern’s losses above $65; if none do, Bankman-Fried will pay Asher the difference between $65 and the maximum loss.” This is now corrected on the web and the Bloomberg Terminal, but if you read Money Stuff by email you got the wrong description of the bet, and that is annoying and I’m sorry. Second, a reader tweeted a better trade: “SBF also either missed or passed on a one-sided bet with greater +EV that would have inflicted greater loss on the counterparty on the first flip: pay two other interns $1 each to coin flip each other for $100.” Yes. This trade would cost Asher $35 (one intern loses $100 on the coin flip, $35 more than the $65 strike price on Bankman-Fried’s bet with Asher[3] ) and would make Bankman-Fried $33 ($35 from Asher minus the $2 he pays the interns). Crucially, Bankman-Fried’s actual trade — pay an intern $1 to flip a coin with him for $98 — had a slightly lower $32 expected value, and was risky; he had a 50/50 chance of winning $130 or losing $66. This trade — pay two interns to flip each other — has a guaranteed profit of $33 with no risk at all. Bankman-Fried went with the riskier trade.